Bits and pieces have been coming in the mail over the last couple of weeks. Finally yesterday the PCBs arrived, so naturally I had to get building.

Unfortunately I hadn't decided on how I was going to reflow the board. I considered one of the cheap death trap $50 hot air stations on eBay, but my bench is small and I'm partial to living in an uncombusted home. I could have picked up a well-regarded (as in: not entirely unsafe) Aoyue station for $100 at the local Fry's, but that was more than I wanted to spend on a merely passable tool. I didn't have time to search thrift shops for the perfect toaster to convert into a reflow oven. I suppose I could have bought a new one, but I get hung up on the allure of reusing an unloved appliance.

Then I stumbled on an old post on piclist mentioning success with butane soldering irons used as small heat guns for SMD work. The catalytic element is the key: it burns the butane at its surface instead of in an open flame, radiating heat while emitting a stream of hot exhaust. The downside is that there's no temperature or airflow control. I figured that I could put a thermocouple probe on the board surface and close the control loop myself.

So that's what I did. I bought a $20 butane iron at Home Depot, along with a cheap ceramic tile as a work surface. The guy at the checkout gave me a funny look, like I was preparing to do... something. Clearly I wasn't remodeling my kitchen or installing a new thermostat. Maybe I was one item short of getting on a watch list.

Anyway, the board soldering was a big success considering it was my first time using stencils and reflow. I set up the board in the jig and stencil, pasted it up, and taped it to the tile. Component placement was fairly easy with some new curved tweezers I bought along with the paste. Unfortunately I fumbled and lost the $0.80 LED, so I left that out for the moment.

I taped the thermocouple to the tile with the tip touching the PCB and fired up the torch. I started slowly, with the torch near its lowest setting, and constantly moving the air stream around the board. I used the torch distance as fine control over the temperature, and when it looked like the board was thoroughly soaked around 160C, I cranked up the torch and watched for the solder to melt while trying to keep the temperature around 240C for about a minute. After everything looked smooth and silvery, I gradually dialed back the torch in a lame attempt to gradually cool down the board.

The result was a reflowed QFN and four parts stuck down to the board on their pads. But the QFN had a solder blob between two pins, and one of the bypass caps wasn't making contact with a pad. The QFN blob was easily fixed with wick, and the cap replaced with an iron. The QFN was probably due to too much paste, so I'll work on my technique next time. The cap however was a silk problem: it looks like the silk was off slightly for the entire backside, and a marker in silk between the pads had partially covered one of the landings.

I found my missing LED stuck to some of the tape I used to secure the board, and added that as well as the 8 pin connector. Turns out that a 1.6mm thick board is way too thin to edge mount a dual row header, so I had to solder-blob one row of pins to the pads. I'm not concerned though, since this header is only for debugging.

I did basic continuity checks as much as I could and cleaned off flux from the hand soldering. By this time it was fairly late and I had to get up early for work today. No time for "Hello World!" but there is that USB bootloader.

I plugged it into my Linux PC and it enumerated on the first try!

Next step: blink an LED.